The ISAD (Infraslow Neurofeedback for Anxiety and Depression) Study: A Protocol for a TransDiagnostic, Randomized, Double-Blind, Sham-Controlled, Dose-Response, Parallel-Group Trial

Background: The core intrinsic connectivity networks (ICNs), encompassing the default-mode network (DMN), salience network (SN) and central executive network (CEN), have been shown to be dysfunctional in individual’s with internalizing disorders (IDs; e.g. major depressive disorder, MDD; generalized anxiety disorder, GAD; social anxiety disorder, SOC). As such, source-localised electroencephalogram neurofeedback (EEG-NFB) therapy targeting key cortical nodes within these networks has the potential to reduce symptoms associated with IDs and restore normal core ICN function. We intend to conduct a transdiagnostic, randomized, double-blind, sham-controlled, dose-response, parallel-group trial of standardized low-resolution electromagnetic tomography electrophysiological infraslow (<0.1 Hz) uctuation neurofeedback (sLORETA eISF-NFB) 3 times per week over 4 weeks in participants (n=60) with IDs. Our primary aim will be to assess the clinical ecacy of sLORETA eISF-NFB via relevant patient-reported outcomes (PROs). Methods: We will randomly assign participants with a current diagnosis of MDD, GAD, and/or SOC to one of four groups: 1) 12 sessions of posterior cingulate cortex (PCC) up-training (n=15), 2) 6 sessions of yoked-sham training followed by 6 sessions of PCC up-training (n=15), 3) 12 sessions of concurrent midcingulate (MCC) down-training and PCC up-training (n=15), or 4) 6 sessions of yoked-sham training followed by 6 sessions of concurrent MCC down-training and PCC up-training. Transdiagnostic PROs, as well as resting-state neuro-physiological measures (EEG; electrocardiography, ECG; electrodermal activity, EDA), will be collected from all subjects at baseline, mid-training, 1 week post-training, and 1 month post-training. We will further compare baseline PROs and neuro-physiological measures to age- and sex-matched non-ID (i.e. no ID The ISAD (InfraSlow neurofeedback for Anxiety and Depression) study is a trans-diagnostic, randomized, double-blind (subjects and raters), sham-controlled (yoked by sex and ID diagnosis), dose-response (6 versus 12 genuine treatment sessions), parallel-group (4-arm) superiority, pilot trial using a 1:1:1:1 allocation ratio with primary endpoints at mid-training (i.e. after 6 sessions), 1-week post-training and 1-month post-training. Importantly, our trans-diagnostic approach heeds recent calls for a more pragmatic, ecologically-valid clinical research(196-206).


Introduction
Background and rationale {6a} Mental disorders are one of the most common causes of morbidity and mortality worldwide(1) with rates markedly increasing in recent years (2)(3)(4)(5)(6). Here in New Zealand, it is estimated that one in ve is dealing with a mental illness at any given time with a majority likely to experience at least one episode at some point in their lifetime (7). Internalising disorders (IDs), which include all anxiety (e.g. generalised anxiety disorder, GAD; social phobia, SOC) and depressive (e.g. major depressive disorder, MDD) disorders, are the most common mental illnesses characterised by neuroticism and distress experienced inwardly(8, 9).
EEG-learning will be evident following active, but not sham, sLORETA eISF-NFB

Pregnant
. Pacemaker Who will take informed consent? {26b} At the initial meeting, a trained male doctoral student will 1) provide each potential participant with a paper copy of the participant information sheet written in English, 2) query if they have read and understood the document, 3) ask if they have any questions about the trial, and 4) request written informed consent from individuals willing to participate in the study. Participants will be informed that they may withdraw at any time without giving a reason and that all data collected up to the point of withdrawal may be used in the nal analyses.
Additional consent provisions for collection and use of participant data and biological specimens {26b} Not applicable: no biological specimens will be collected and all data is to be used solely in accordance with this trial.

Interventions Explanation for the choice of comparators {6b}
The non-ID control group, which does not receive any intervention, will only be assessed at baseline in order to ascertain differences from the ID population in baseline PROs and neuro-physiological measures. To mitigate against selection bias (241), the same inclusion and exclusion criteria apply to the non-ID controls with the exception an ID diagnosis (e.g. MDD, GAD, SOC).
In the clinical cohort, our choice of sham-controls allows us to elucidate any potential speci c (e.g. nonplacebo) effects and addresses widespread concerns of generally weak methodological designs in NFB trials(35, [242][243][244]. Further, the staggered start allows us to perform a dose-response assessment and assures all clinical participants receive some genuine sLORETA eISF-NFB thereby minimizing the potential of sham-trial associated participation barriers (245) as well as addressing any potential ethical concerns of sham-only allocations.

Intervention descriptions {11a}
Following baseline assessments, clinical participants will be randomised to one of 4-arms: 1) 12 sessions of PCC eISF up-training, 2) 6 sessions of yoked-sham followed by 6 sessions of PCC eISF up-training, 3) 12 sessions of concurrent PCC eISF up-training and MCC eISF down-training, 4) 6 sessions of yokedsham followed by 6 sessions of concurrent PCC eISF up-training and MCC eISF down-training. Training will commence within 1-week of baseline #2.
To help reduce impedances, subjects were asked to arrive with non-braided, dry, clean (i.e. no conditioner, gels, pastes, sprays) hair. Participants will attend three 30-minute sessions per week, every other day, over 4 weeks (12 sessions in total). 19-channel sLORETA eISF-NFB training will be performed using a 21channel DC coupled ampli er produced by Brainmaster Inc. and the BrainAvatar software (version 4.7.5.844) in a quiet, cool (~15°C), dimly lit room by an unblinded male researcher with >2 years of experience in the administration of NFB. Participants will be seated in a comfortable chair and an appropriately sized Comby EEG cap will be placed on the participant's head. Using a blunt need and syringe, the scalp will be mildly abraded prior to the application of an electrolyte gel beneath each electrode. EEGs will be recorded with the silver/silver chloride (Ag/AgCl) electrodes positioned according to the International 10-20 system (i.e., Fp1 , Fp2, F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T3, T4, T5, T6, Fz, Cz, Pz) using a linked mastoids reference and a ground electrode positioned centrally between, F3, Fp1, Fz and Fpz. It should be noted that the purpose of the cool room and scalp abrasion is to mitigate contamination of the EEG signal by electro-dermal (i.e. sweat gland) potentials which are known to mimic brain-derived ISFs (115,121). The impedances of the active electrodes will be kept below 10 k Ω and a 50 Hz notch lter will be set. During the set-up, participants will be asked to complete the DESS on an IPad.
Immediately prior to each training period, a demonstration of motion/electromyography (EMG) artefact alerts will be performed with instructions to avoid eye/head/face movements in order to minimize this non-rewarding feedback. Participants will then be instructed to close their eyes, relax, stay awake, and listen to the sound being played. They will be informed that the sound they hear re ects that they are doing well. Notably, no explicit strategies or instructions were given as, with few exceptions(246), implicit strategies have been shown to produce better outcomes (247)(248)(249)(250)(251)(252).
Continuous, real-time auditory feedback will be used for reinforcement and produced within 30 milliseconds of the subject's ISFs (0.0-0.1 Hz) within the pre-de ned region(s) of interest (i.e. MCC and/or PCC) surpassing the threshold(s). These ROI were selected because, as outlined in the introduction, they are considered key cortical nodes within the core ICNs which are consistently found to be disrupted in ID populations. sLORETA permits the selection of any cortical region for feedback of the current density using voxels selected based on Montreal Neurological Institute (MNI)-coordinates (253). For a complete list of targeted voxels for this trial, see Additional les 1 and 2.
The reward threshold(s) will be manually adjusted in real-time to maintain a 60% ± 10% success rate.
Manual, rather than automated, thresholding was chosen as it has been reported to lead to better EEGlearning (243,248,250,254). The yoked-sham sessions will be identical to active sessions, including live EEG recordings and real-time motion/EMG artefact alerts, however the auditory rewards will derive from playbacks of a series of 6 consecutive ISF-NFB sessions from another female with IDs recorded via free, open-source Audacity software(255) which uses the computer's sound card as an audio to digital converter. Importantly, it has been reported that training effects are more robust when the clinician is present(256) therefore, irrespective of group assignment, the trainer was present for the duration of all sessions. Further, the trainer will monitor subject protocol adherence.
Criteria for discontinuing or modifying allocated interventions {11b} Participants may withdraw at any time without giving a reason or may be withdrawn by the investigators if they are unable to adhere to protocol (i.e. miss >1 eISF-NFB session), start or alter other therapies (e.g. start/change medications), or experience signi cant adverse effects.
Strategies to improve adherence to interventions {11c} We will attempt to mitigate adherence issues via automated email and text message reminders sent on the day of each training session.

Relevant concomitant care permitted or prohibited during the trial {11d}
Participants should continue to take medications for other conditions as normal and will be asked to maintain any current rst-line mental health therapies (e.g. pharmacotherapy) for the entire length of the study period. Any changes to these rst-line therapies (e.g. altered pharmacotherapy dosages, introduction of intensive psychotherapy) will render participants ineligible.
See Table 1 for a detailed description of the trial intervention using the Template for Intervention Description and Replication (TIDieR, (257)) Provisions for post-trial care {30} In the unlikely event of injury, participants will be eligible to apply for compensation from the Accident Compensation Corporation (ACC) of New Zealand just as they would be if they were injured in an accident at work or at home. Although there are private providers abroad (e.g. Asia, North America, Europe), should this trial provide evidence of e cacy, there is currently no access to this therapy within New Zealand.
Outcome measures will be assessed at baseline on two separate occasions approximately 1-week apart followed by mid-training (i.e. between the 6 th and 7 th ISF-NFB session), 1-week post-training, and 1-month post-training assessments. The duplicate baseline assessments allow us to examine state-variability and test-retest reliability. The mid-training assessments permit comparisons of active vs sham (i.e. active PCC vs sham PCC and active PCC/MCC vs sham PCC/MCC) and single-vs multi-region (i.e. active PCC vs active PCC/MCC) groups. The post 1-week and post 1-month follow-ups allow for the determination of any immediate and delayed treatment effects, respectively, following either 6 or 12 active sessions. For all outcomes, the analysis metric of interest will be group mean (or median) change from baseline.

Sample Size {14}
Our target sample sizes for our clinical and healthy populations are n=60 and n=60, respectively, which is based on trial feasibility during the time of recruitment. Based on our prior feasibility study (194), the projected rate of loss to follow-up/discontinuation following randomisation is 10-15%. To date, this is the rst study examining the effect of sLORETA eISF-NFB on IDs, thus no formal power calculations can be made. Our research group's previous sLORETA eISF-NFB feasibility study found signi cant differences in PROs and neurophysiological measures between the active and sham groups with an n=11 and n=10, respectively (194,195).

Recruitment {15}
A convenience sample will be recruited will be via posters and Facebook ads with an invitation to participate in a University of Otago mental health study. Advertisements will direct potential participants to a webpage that will describe the trial and invite those interested to complete an online screening form which will query basic information including rst name, age, date of birth, sex, ethnicity, education level, handedness, current diagnoses and medications, pregnancy status, presence of electronic implants (i.e. pacemakers), email address, and phone number. Individuals who pass the preliminary screening will be contacted via email and asked to attend an in-person mental health interview at the University of Otago Hospital, Dunedin, New Zealand. Those that agree will be provided directions to the lab and a digital copy of the 6-7 page participant information sheet. A reminder text will be sent to potential participants on the day of their interview. Each person will be screened using the Mini-International Neuropsychiatric Interview (MINI; English version 7.0.2 for DSM-5)(268) by a extensively trained doctoral student. The MINI is a brief structured diagnostic interview, shown to be both valid and reliable, used to assess the 17 most common psychiatric disorders including MDD, suicidality, bipolar, panic disorder, agoraphobia, SOC, obsessive compulsive disorder, post-traumatic stress disorder, alcohol use disorder, substance use disorder, psychoses, anorexia, bulimia, binge-eating disorder, GAD, and anti-social personality disorder(269, 270). In the event that the interviewer suspects that the interviewee is at high-risk for suicide, he will screen using the Columbia-Suicide Severity Rating Scale (C-SSRS -Screen Version; (271)) with a rmative answers to questions 4, 5, and/or 6b initiating immediate referral to Emergency Psychiatric Services. Ultimately, those meet the inclusion criteria and none of the exclusion criteria will be enrolled into the study, have anthropometric (i.e. height and weight) measurements taken, and scheduled for their baseline assessments. Participants will also be familiarized with the study equipment, procedures, and personnel. Recruitment will continue until our target sample sizes (n=60 clinical and n=60 healthy) is met and is expected to take up to 12 months. All participants who complete the study will receive a $40 supermarket voucher as reimbursement for any parking/travel expenses.

Sequence generation {16a}
Randomisation will be performed using the website randomization.com. This tool is a valid randomisation program utilised by clinical trial researchers. To ensure balanced sample size across the groups over time, block randomisation with random block sizes and a 1:1:1:1 allocation will be utilised.

Concealment mechanism {16b}
To ensure concealment, the block sizes will not be disclosed. Further, treatment conditions will be kept in sequentially number, sealed, opaque envelopes in the central o ce.

Implementation {16c}
A lab member from the group who has no direct contact with the participants will conduct the randomisation process. TMP is responsible for participant enrolment and will assign participants to interventions following baseline assessments and con rmation of scheduled eISF-NFB sessions.

Who will be blinded {17a}
This is a double-blind study whereby trial subjects and outcome assessors will be unaware of group assignments. In an effort to improve blinding, all aspects of sham sessions will be identical to active sessions including the live recording of sham participants' EEGs along with real-time artefact alerts.
Blinding integrity will be assessed during the mid-training assessment session via an IPad whereby participants will be queried as to 1) their perceived group allocation, 2) con dence in their answer to question 1 on a scale of 0-100%, 3) reason for their answer to question 1, and 4) if their group assignment was revealed to them in any way.

Procedure for unblinding if needed {17b}
Treatment assignment will be disclosed to trial participants only upon their completion of the study.

Data collection and management
Plans for assessment and collection of outcomes {18a} All assessment sessions for a given subject will take place at approximately the same time of day and be led by a female research assistant. Prior to each assessment session, participants will be asked to abstain from 1) food and water for 2 hours, 2) smoking/vaping for 8 hours, 3) strenuous exercise, alcohol, caffeine & over-the-counter medication for 24 hours. A reminder email and text will be sent to each participant one day prior to and on the day of assessment sessions, respectively, in an effort to increase adherence. Adherence to lifestyle restrictions will be queried at the beginning of each session with any breaches recorded. In addition, the subject's previous night's sleep duration will be documented and they will be asked to use the toilet immediately prior to testing to ensure that their bladders are empty. Together, these standardisation procedures will help to control for state-variability stemming from in uences like circadian rhythms (272), satiety (273), sleep duration (55), hydration levels (274), and smoking (275) and are in line with current recommendations for neurophysiological data collection(276).
Further, two baseline assessments will be acquired on each participant approximately 1 week apart in order to assess state-variability and test-retest reliability. Duplicate baseline follows guidance to improve the reliability of brain imaging-related metrics via multiple (i.e. >1) resting-state scans(277-280).
All PROs are English versions and have been created in digital form via Qualtrics(281) which will allow participants to complete them using an IPad during their EEG set-up and prior to all neurophysiological assessments. The order of PRO administration will be standardized and based on PRO length (i.e. IDAS-II > MEDI > HADS > IUS-12 > RTQ-10). To prevent missing data, a visual alert will be generated if any queries on a given form have missing responses. Research has indicated the electronic data collection increases the speed, accuracy, and user acceptability of the process(282-284). The estimated total time to complete the battery of PROs is 20 minutes.

Primary outcomes
The HADS is a valid and reliable 14-item, trans-diagnostic PRO measure used to assess anxiety and depression severities(263). Response options are on a 4-point scale (0-3) based on participants experiences over the past week with anxiety and depression subscale scores graded as follows: 0-7=normal, 8-10=borderline abnormal/borderline case, 11-21=abnormal/case(263, 285). The HADS has been repeatedly shown to be a reliable and valid tool across a variety of settings(285-287). There is some debate with respect to whether the HADS is best assessed via the total 14-item score(288-290) or two 7-item subscale (anxiety and depression) scores(285, 287, 290, 291).
The MEDI is a valid and reliable 49-item, trans-diagnostic PRO measure that employs a response scale ranging from 0 (not characteristic of me/does not apply to me) to 8 (extremely characteristic of me/applies to me very much) to assess nine ID-related symptom domains, originally proposed by Brown and Barlow (293)  for positive temperament), >6 (or <2 for positive temperament), and 7-8 (or 0-1 for positive temperament) may re ect moderate, severe, and extreme severities, respectively (295).

Secondary Outcomes
The IUS-12 is a valid and reliable 12-item, transdiagnostic PRO measure that assesses the degree to which an individual considers the possibility of a negative event occurring unacceptable, irrespective of its probability of occurrence(264, 296). Whereas the original IUS-27 was GAD-speci c, the IUS-12 has been distilled in order to measure the core intolerance of uncertainty construct (297). The IUS-12 uses a response scale from 1 (not at all characteristic of me) to 5 (entirely characteristic of me)(264). Notably, although some researchers have claimed that the IUS-12 is a unidimensional construct and recommend using only the IUS-12 total score(298-300), there has been considerable support for a two-factor IUS-12 structure: 1) a 7-item prospective IU scale related to action/approach-oriented strategies in order to increase certainty (e.g. seeking more information), and 2) a 5-item inhibitory IU associated with inaction/avoidance-oriented thoughts and behaviours (e.g. delayed decision making)(264, 296, 301-308).
Notably, IUS is a common trait shared across the ID spectrum(308-311).
Resting-state full-band EEGs (fb-EEGs), electrocardiograms (ECGs), and EDAs will be collected concurrently with a Compumedics Neuroscan SynAmps RT DC/AC ampli er using a continuous sampling rate of 1000 Hz and 50 Hz notch lter. Recordings will take place in a quiet, cool (~15°C), dimly lit room as participants are seated upright in a comfortable chair with their eyes closed. Notably, EEG reliability appears to be improved in the eyes closed condition(317, 318).
10-minute resting-state fb-EEGs will use high-density (64-channel) silicone Quik-Cap Hydro Net caps with Ag/AgCl electrode placements corresponding to the international extended 10/20 system. The ground electrode is positioned at AFz with the reference electrode midway between Cz and CPz. Electrooculography (EOG) will track vertical and horizontal eye movement artefacts. The cap is soaked in a saline solution at least 30 minutes prior to application and all electrode impedances will be kept below 10 kΩ. To help reduce impedances, subjects were asked to arrive with non-braided, dry, clean (i.e. no conditioner, gels, pastes, sprays) hair. fb-EEG recordings will be for 10-minutes.
10-minute resting-state EDA (in microsiemens, µS) will assessed with bipolar nger Ag/AgCl electrodes securely attached to the palmar surface of the middle phalanxes of the 4 th and 5 th digits (i.e. ring and little ngers) of the non-dominant hand.
Resting-state heart rate will be monitored using a standard limb lead (lead-II) ECG at rest using Ag/AgCl electrodes. A 10-minute spontaneous breathing ECG will be collected concurrently with the other neurophysiological assessments followed immediately by a 10-minute metronome paced breathing (12 breaths per minute) ECG with a 1:1 inspiratory/expiratory (I/E) ratio (i.e. 2.5 sec inhalation/2.5 sec exhalation). Participants will be instructed to breathe through their nose and at normal depth (i.e. no deep breathing).

Plans to promote participant retention and complete follow-up {18b}
Once enrolled, every reasonable effort will be made to follow participants throughout the entirety of the study period via ongoing email and text messaging correspondence. In the event of premature discontinuation of the study for any reason, participants will be made aware that all data collected up to the point of withdrawal may be used for analyses.

Data management {19}
Participant paper les, including case-report-forms and MINI assessments, are to be kept in numerical order and stored in a locked le drawer in the lead researcher's o ce. PROs will be electronically stored in Qualtrics with a back-up copy automatically generated and sent to the lead researcher's trial email address. All data collected will be entered into Microsoft Excel (2016) and double-checked for accuracy by the data analyst at the time of entry. Participant data will be maintained for a period of not less than 10 years after the completion of the study.

Con dentiality {27}
All information generated in this study will be considered highly con dential and is not to be shared with any persons not directly concerned with the study. For de-identi cation purposes, participants will be assigned unique study numbers upon enrolment (e.g. ISAD_F8 = 8 th enrolled female in the ISAD trial). All electronic records will be identi ed solely using assigned study numbers and stored locally in a password protected database. All paper records will be stored on-site in a locked o ce accessible only to the researchers directly involved in the trial. Further, paper documents that contain personal identi ers (i.e. informed consent forms), will be store separately from de-identi ed paper records (i.e. case-report-forms and MINI assessments).
Plans for collection, laboratory evaluation and storage of biological specimens for genetic or molecular analysis in this trial/future use {33}an Not applicable: no specimens collected Statistical methods Statistical methods for primary and secondary outcomes {20a} SPSS (version 27) will be used for PRO statistical analyses. PRO group means/standard deviations will be calculated at each time point. Baseline #1 and #2 values will be used to assess test-retest reliability for all PROs in both the non-ID and ID cohorts. An independent t-test (95% con dence intervals) will be used to compare baseline #2 values of the non-ID and ID cohorts.
In our ID cohort, change from baseline #2 values will be statistically compared using repeated measures ANOVA (95% con dence intervals) with group allocation as the between-subjects factor and time as a within-subjects factor. Change from baseline will be considered at three time-points: 1) mid-training (i.e. between session 6 and 7), 2) 1-week post-training, and 3) 1-month post-training. Due to the large number of comparison, appropriate controls for multiple comparisons will be applied. In addition, if feasible, we will undertake Bayesian analyses in R (version 4.0.2; (319)) with posterior distributions and corresponding 95% credibility intervals reported.
For fb-EEGs, raw data will be resampled at 128 Hz, ltered from 0.001 Hz to 44 Hz, plotted, and carefully inspected using manual and/or automated artefact rejection in EEGLAB (version 2020). Subsequently, average cross-spectral matrices will be computed for bands infraslow (0.01-0.1 Hz), slow (0.1-1 Hz), delta (1-4 Hz), theta (4-8 Hz), alpha (8-12 Hz), beta (12-30 Hz) and gamma (30)(31)(32)(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43)(44). A validated sourcelocalization technique, exact LORETA (eLORETA)(320, 321), will be used to in order to determine intracerebral electrical sources of the scalp-recorded EEG activity in each of the frequency bands. Using the Montreal Neurological Institute (MNI)-152 template, eLORETA produces an inverse solution space consisting of 6239 cortical grey matter voxels at 5mm resolution and has been shown to produce exact, zero-error localisations even in the presence of structured noise (321). The eLORETA software package performs the statistical analyses using a non-parametric methodology. It is based on estimating, via randomization, the empirical probability distribution for the max-statistic (e.g. the maximum of a t or an F statistic), under the null hypothesis while correcting for multiple testing (i.e., for the collection of tests performed for all electrodes and/or voxels, and for all time samples and/or discrete frequencies). The interested reader is referred to (322) for a complete overview of the methodology, where details about the properties (e.g. pertaining to its non-parametric nature, and pertaining to how it properly corrects for multiple testing) can be found The raw ECG signal will be band-passed, differentiated, squared, integrated, and smoothed to isolate Rwaves. Manual inspection of the ECG trace will be performed for artefact (e.g. ectopic beats) identi cation and correction. Short-term (5-10 min) HRV using both spontaneous-and paced-breathing ECGs will be calculated using standard (low = 0.04-0.

Interim analyses {21b}
Not applicable: no interim analyses will be performed and no stopping guidelines will be established.

Methods for additional analyses (e.g. subgroup analyses) {20b}
If primary outcomes do not signi cantly differ between active groups at each time point, combined between-group comparisons (active PCC & PCC/MCC vs sham>active PCC & PCC/MCC) will also be performed using the methods outlined above.
Methods in analysis to handle protocol non-adherence and any statistical methods to handle missing data {20c} For data to be included in the current study, participants had to schedule and complete a minimum of 10 out of 12 eISF-NFB sessions. Modern imputation methods will be applied for missing data re ecting uncertainty in the modelling process and inherent variability in patient outcomes. Further, we will report and qualitatively compare reasons for withdrawal in each of the randomisation groups. Based on our prior feasibility study (194), discontinuation/loss-to-follow-up during the intervention/post-intervention phases is expected to be 10-15%.
Plans to give access to the full protocol, participant level-data and statistical code {31c} The full protocol will be published in a peer-reviewed, open-source journal prior to recruitment completion.
No more than 2 years following the nal data collection, we will deliver the completed, de-identi ed dataset and statistical code to the appropriate data archive for sharing purposes in line with the scienti c imperatives of increased transparency, reproducibility, and interpretation of trials.

Oversight and monitoring
Composition of the coordinating centre and trial steering committee {5d} Not applicable: no coordinating centre or trial steering committee for this trial Composition of the data monitoring committee, its role and reporting structure {21a} Due to the relatively short duration of recruitment, non-invasive make-up of the procedures/interventions, and non-serious nature of adverse effects reported in our prior feasibility trial, no formal data monitoring committee will be established.

Adverse event reporting and harms {22}
We will systematically monitor adverse effects from the therapy for the duration of the trial using the Discontinuation-Emergent Signs and Symptoms checklist (DESS; (327)) created, verbatim, in Qualtrics and completed by participants on an iPad during EEG set-ups in the interventional and post-interventional phases. Initially developed for drug trials (327), the DESS is a structured 43-item self-report that utilises the following scale: 1=new symptom, 2=old symptom but worse, 3=old symptom but improved, 4=old symptom but unchanged, 5=symptom not present. The DESS has been used for the assessment of treatment-related side-effects in ID populations(328, 329) and, recently, has been employed to monitor adverse-effects speci cally associated with NFB therapy (330). Participants may be withdrawn from the trial by the investigators, even without their request, in the event of serious adverse effects. As detailed in the PIS, a brief (4-item) interview used during our group's prior sLORETA eISF-NFB feasibility trial revealed that, although unusual or vivid dreams were experienced by some participants, there were no serious adverse effects (195).

Frequency and plans for auditing trial conduct {23}
Not applicable: no auditing of trial conduct will be performed.
Plans for communicating important protocol amendments to relevant parties (e.g. trial participants, ethical committees) {25} Substantive protocol amendments which may impact on the conduct of the study including changes to the study objectives, design, population, sample sizes, or procedures will be agreed upon by the research team, updated in the trial registry, submitted to the ethics committee for approval and updated on our online trial advertisements and web pages.

Dissemination plans {31a}
Every effort will be made to minimise the interval between the completion of data collection and release of study results. We estimate this process to take 6-12 months. Irrespective of magnitude or direction of effect, results from the study will be written up and submitted to international peer-reviewed scienti c journals, presented at scienti c conferences, and may form part of grant applications. In addition, once compiled, all participants will be provided with a digital copy of the results.
[1] The criterion does not apply to non-ID controls.
Page 18/49 [2] The non-ID group will undergo the MINI and baseline assessments only.

Discussion
Approximately one in ve New Zealander's is dealing with a mental illness at any given time with the majority experiencing psychopathology at some point in their lifetime (7). Alarmingly, New Zealand's suicide numbers are increasing with the 2017-18 rate the highest it's been in 20 years (7).
Psychopathology has been estimated to reduce life expectancy by up to 25 years (7). A recent government inquiry by the New Zealand government has shed light on the shortcomings of current treatment and called for wider implementation of non-pharmaceutical approaches in treatment of mental health problems (7). Similarly, scientists in other parts of the world are calling for research into 'novel interventions that may be based on altering plasticity or returning circuitry rather than neurotransmitter pharmacology' (331).
The implementation of safe, non-invasive neuromodulation techniques that have the potential to impact neuroplasticity within and between large-scale ICNs may offer new treatment opportunities for individuals who either do not want, respond to, or tolerate standard interventions. Additionally, these techniques may serve as adjuncts to traditional treatments, potentially enhancing their e cacy. To date, ours is the only research group studying the effects of this novel therapy in clinical populations. We believe targeting core We are grateful to all of the trial participants for their willingness and commitment to participate in this trial.

Authors' contributions {31b}
TMP conceived of the study and drafted the protocol. TMP, PG, DDR, and DA contributed to the re nement of the trial protocol. TMP will conduct the trial, statistical analyses and dissemination of results. PG provided expertise on mental health disorders. DDR provided neuroscienti c expertise. DA provided statistical expertise. MS created the neurofeedback software programs and provided training in their implementation. All authors read, provided feedback, and approved the nal manuscript.

Funding {4}
This clinical trial is part of a doctoral thesis supported by the Department of Surgical Sciences, University of Otago, Dunedin, New Zealand.

Availability of data and materials {29}
The nal trial dataset will be password protected and housed locally at the research lab. Other team members will be provided access to this dataset by TMP upon request. To ensure con dentiality, data dispersed to project team members will be blinded of any identifying participant information. Written, informed consent to participate will be obtained from all participants.

Consent for publication {32}
Examples of the participant-information and informed-consent forms can be provided upon request.

Competing interests {28}
MS is the owner of Neurofeedback Therapy Services of New York which provides ISF-NFB therapy for patients and training for clinicians. He created the ISF-NFB software programs for this trial, provided guidance to the principal investigator in their utilisation, and participated in editing of the nal version of the manuscript. MS did not participate in the trial conceptualisation, design, implementation, or analyses. All other authors declare no competing interests.  eISFs are believed to coordinate and integrate information exchange within and between core-ICNs.
Further, communication within and between core-ICNs has been found to be disrupted in ID populations. We hypothesize that eISF-NFB targeting key cortical nodes of these core-ICNs, may restore proper intra-and internetwork function and reduce ID-related symptoms.  the head and, using a blunt need and syringe, the scalp will be mildly abraded just prior to the application of the electrolyte gel beneath each electrode. EEGs will be recorded with the Ag/AgCl electrodes positioned according to the International 10-20 system (i.e., Fp1 ,   Fp2, F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T3, T4, T5,   T6, Fz, Cz, Pz) using a linked mastoids reference and a ground electrode positioned centrally between, F3, Fp1, Fz and Fpz. Impedances will be kept below 10 kΩ and a 50 Hz notice filter will be set.
Immediately prior to each training period, a demonstration of motion/EMG artefact alerts will be provided with instructions to avoid eye/head/face movements in order to minimize this non-rewarding feedback. Participants will then be instructed to close their eyes, relax, stay awake, and listen to the sound being played. They will be informed that the sound they hear reflects that they are doing well.
Continuous, real-time auditory feedback (organ tones) will be used for reinforcement when the subject's eISFs surpass the threshold(s). The reward threshold will be manually adjusted in real-time to maintain a 60% ± 10% success rate. The yoked-sham sessions will be identical to active sessions, including live EEG recordings and realtime motion/EMG artefact alerts, however the auditory rewards will derive from playbacks of consecutive, prerecorded sessions of another female with IDs.
The trainer will remain present for the duration of all sessions to monitor the EEG.  Figure 1